Can You Use an EA at a Prop Firm? Rules and Limits

Written by Dominic Walsh · Published 12 Aug 2026 · Last updated 17 Aug 2026

Can you use an EA at a prop firm? In most cases yes, and the qualifier carries more weight than the answer itself. Every firm writes its own rulebook, and that rulebook decides which robots survive a challenge.

So the useful question changes shape. Ask which categories of restriction apply, then ask whether your program obeys them without anyone watching the screen at three in the morning.

Can You Use an EA at a Prop Firm? The Short Answer

Most funded programmes permit automation of some kind. A few allow any robot you like, several allow only code you wrote yourself, and a small number ban automation outright.

Rules also move. A firm can rewrite its terms between one challenge and the next, so a summary you read last year may no longer describe the agreement you signed today.

The panel above sets out four checks that decide the matter. Each row either clears your program or blocks it, and one blocked row ends the account.

Why the Answer Varies So Much

Prop firms carry the risk behind your trades, either on their own book or through a hedge. Their rules cap that risk, so they rarely exist to make trading harder for its own sake.

A robot that trades once a day looks harmless from their side. One that fires forty orders a minute around a data release looks like a cost, and they price it as one.

Because of that, two firms can reach opposite conclusions about the same program. Neither firm behaves unreasonably; they simply hedge differently.

Read the Rulebook, Not the Marketing

Terms live in the account agreement rather than on the landing page. Open that document, search for “expert advisor”, “automation”, “algorithm” and “copy”, then read every hit slowly.

Ask support in writing whenever wording reads two ways. A saved email settles a dispute far better than a memory of a chat window.

Our overview of prop firm EA rules maps the common ground. Use it as a map, then verify every point against the terms you actually accepted.

The Rule Categories That Decide It

Restrictions fall into a handful of families. Learn the families, and any new rulebook then reads in about ten minutes.

The diagram above walks through them in the order that matters. Loss caps come first, because they end accounts faster than anything else on the list.

Maximum Daily Loss

This cap limits how far an account may fall within one trading day. Breach it and the account closes, whatever price does afterwards.

Most firms measure the cap against equity rather than balance. So an open losing position counts against you long before you close it.

Maximum Overall Loss

The overall cap works the same way across the whole life of the account. Some firms fix it to the starting balance, while others trail it behind your highest equity.

Trailing versions bite hardest. After a good run the floor sits above your starting point, so a modest giveback can end an account that still shows a profit.

News Windows

Many firms restrict trading around high impact releases. The window usually runs a few minutes either side of the print.

Some ban new entries only. Others ban holding any position through the window, which changes what your program must do rather than merely when it may act.

Copy and Duplicate Trading Bans

Firms often forbid mirroring identical trades across several funded accounts. They also forbid pulling another trader’s orders into a funded account through a bridge.

Hedging explains the rule. Identical fills across many accounts turn one idea into one very large exposure on their side of the book.

High Frequency and Latency Restrictions

Rules here target extremely short holding times, tick scalping and any attempt to profit from a stale quote. Thresholds vary, and some firms state one in seconds.

A grid program or a news scalper can trip this clause without intending to. Check your average holding time before you assume the rule leaves you clear.

Consistency Rules

A consistency rule caps how much of your total profit may come from a single day or a single trade. It filters out the lucky one-off result.

Wording matters enormously here. Some firms apply the test only at payout, while others apply it throughout, and our note on the prop firm consistency rule shows how the arithmetic lands.

How a Daily Loss Limit Ends an Account

This rule catches more automated accounts than any other. Its mechanics deserve a close look.

The panel above shows the failure exactly. Equity dips through the limit during the session, the breach registers at that moment, and the account closes even though the day finishes back above the line.

Why the Cap Reads Equity, Not Balance

Balance ignores open trades. Equity does not, so a floating loss can breach the cap while every position remains open and every stop remains untouched.

Code the check against equity for that reason. A robot watching balance alone will miss the only number the firm cares about.

The Reset Clock

Daily caps reset at a stated server time, often the broker’s midnight. That moment rarely lines up with your own.

A program holding trades across the reset can open a fresh day already deep in a floating loss. Find the server hour first, then build any schedule around it.

Sizing Backwards From the Cap

Work from the cap towards the lot size, never the other way round. If your program may hold three positions at once, three simultaneous stops must land well inside the daily limit.

Run that arithmetic through our prop firm position size calculator. Seeing the worst case in lots tends to change how the settings look.

Building an EA That Obeys the Rules

A rulebook helps only once the code enforces it. Every restriction needs a matching guard inside the program.

The comparison above pairs each written rule with the code that has to carry it. Any rule without a guard depends on you noticing in time, which rarely happens overnight.

Hard Guards Beat Good Intentions

Put a daily drawdown check at the top of the trading logic. Once the day’s loss passes your own softer threshold, the program stops opening trades until the reset.

Set that threshold below the firm’s number. A margin of roughly a third leaves room for slippage on the way out.

Time and News Filters

Add a session filter and a news filter. Both take a few lines of code, and both prevent the two most common breaches.

Feed the news filter from a list you control. Our economic calendar shows the releases that most rulebooks restrict.

One Account, One Instance

Run a single instance per account, each with its own magic number. Duplicate instances create duplicate orders, and duplicate orders look exactly like copy trading from outside.

Log every order the program sends. If compliance ever queries your activity, a log answers in minutes, while memory answers badly.

Inputs deserve their own pass too, and our guide to expert advisor risk settings covers the ones that matter most.

News Windows and Trading Blackouts

News rules confuse more traders than any other category. Wording varies between firms, and the consequences do not.

The timeline above shows the shape of the problem. Two orders land inside the shaded window, and both fall foul of the rule whatever their outcome.

What Counts as Restricted News

Firms usually name a tier of releases: rate decisions, employment data and inflation prints. Some publish an exact list, while others point at a third party calendar and leave you to match it.

Build your filter from their list rather than your own view of what moves price. Their list decides the breach, so their list wins.

Entries, Exits and Holding

Check whether the rule bans entries, exits or both. A ban on holding through the window forces the program to flatten before the window opens.

That changes the strategy rather than the settings. A swing rule may not survive a requirement to close everything twice a week.

The Practical Fix

Stop trading a few minutes earlier than required, then resume a few minutes later. Server clocks drift, and the buffer costs very little.

Keep the buffer as an input rather than a constant. Rules shift, and an input takes ten seconds to change.

What Happens to Open Trades

Decide in advance what the program does with a position already open when a window arrives. Three options exist, and each carries a different cost.

Closing early gives up part of the move. Moving the stop to breakeven protects the account without a full exit, though a fast print can jump straight past it.

Leaving the trade alone breaches the stricter rulebooks. So the choice depends entirely on the wording, which brings you back to reading the agreement carefully.

Weekend and Rollover Edges

Some firms treat the Friday close and the Sunday open as restricted too. Gaps hurt their hedge in the same way a data print does.

Add a weekend rule to the program even when the terms stay silent. A gap through your stop costs more than a missed Monday entry.

What Compliance Teams Actually Watch

Reviews happen on flagged accounts rather than on every account. Certain patterns raise the flag.

Identical Fills Across Accounts

Two accounts filling the same instrument inside the same second look like one strategy split in two. That pattern draws attention even when the accounts belong to different people.

Very Short Holding Times

A cluster of trades lasting a second or two suggests latency arbitrage. Firms read that as an attack on their pricing rather than as trading.

One Trade Carrying Everything

A single trade producing most of the profit trips consistency reviews. It also tells the firm almost nothing about whether you can repeat the result.

Orders Clustered on the Release

Entries that land in the same second as a data print look like news arbitrage. Firms watch for that timing because their hedge fills worst at exactly that moment.

A slow program rarely triggers it. One that polls for a price change every tick can, even when the strategy behind it looks ordinary.

Sudden Size Changes

A jump from small lots to very large ones late in a challenge reads as gambling rather than as a system. Some rulebooks call that out directly.

Keep sizing proportional throughout. A steady formula also makes the record easier to defend if anyone asks.

Copying somebody else’s orders raises the same flag from another direction, and our guide to copy trading explains why the ban exists at all.

Settings That Suit a Funded Account

A robot tuned for your own account rarely suits a funded one. Three settings usually need changing before the first trade.

Risk Per Trade

Cut it, often by half. A challenge rewards survival far more than speed, and the loss caps punish one bad week harshly.

Smaller size also buys time. More trades then fit inside the same cap, so the sample grows before any limit gets close.

Maximum Open Positions

Cap simultaneous trades, then cap correlated ones separately. Three long positions in related pairs behave as one larger trade when news lands.

Trade Frequency

Slow the program down when its holding time sits anywhere near a stated threshold. Fewer and longer trades keep you clear of the high frequency clauses.

Signal quality still matters, of course, and our pillar guide to MT4 indicators covers the tools these programs read. Firm rules apply whatever the source of the signal.

Who Owns the Code Also Matters

Several firms ask where your program came from. The question sounds administrative, yet it decides whether a breach becomes a dispute.

Bought, Rented or Written

A program you wrote yourself raises no ownership question. One you bought usually needs a licence that permits use on a funded account, and vendors differ on that point.

Rented programs cause the most trouble. If the licence expires mid-challenge, the terminal keeps running while the logic quietly stops placing orders.

Shared Programs Look Like Copying

A popular commercial robot places similar orders for everybody running it. Firms notice that clustering, and some restrict widely sold programs for exactly that reason.

Ask before you pay a fee. Support can usually tell you whether a specific vendor already sits on a restricted list.

Keep the Paperwork

Save the licence, the purchase record and the settings file you ran. Compliance questions arrive weeks later, and reconstructing that evidence afterwards wastes days.

Challenge Phase Against Funded Phase

Many traders assume one rulebook covers both stages. Several firms tighten or loosen terms once the account turns funded.

Targets Disappear, Caps Remain

A profit target usually applies only during the challenge. Loss caps carry straight through, and they matter more once real payouts depend on them.

Payout Conditions Arrive

Funded accounts add rules about minimum trading days, withdrawal timing and sometimes consistency. Read those clauses before the challenge, not after passing it.

Scaling Changes the Arithmetic

Some firms raise the account size after a good period. Your program then trades a larger balance under the same percentage caps, so absolute risk grows while the settings look unchanged.

Check the sizing formula at every scale step. A fixed lot input turns dangerous the moment the balance moves.

Testing an EA Against Prop Rules

Testing for a funded account differs from ordinary testing. You test survival against a rulebook rather than profit alone.

Replay the Rules Over the Backtest

Export the trade list, then walk the equity curve day by day. Mark any day where the intraday drawdown would have crossed the cap.

Most people find at least one. That single day would have ended the account, whatever the final total shows.

Use the Worst Stretch, Not the Average

Averages hide the day that matters. Find the deepest run of consecutive losses, then measure it against both caps.

Our note on drawdown in trading explains why depth and duration deserve separate attention.

Forward Test on a Demo Challenge

Run the program on a demo account with the same caps for several weeks. Mechanical errors surface quickly, and none of them cost anything there.

Only then commit a fee. A demo failure costs a fortnight; a live failure costs the challenge.

Why Automated Funded Accounts End

Failures repeat themselves. Five causes cover most of them, and none involves a clever rule that suddenly stopped working.

1. No equity guard. The program watched balance, so a floating loss breached the daily cap unnoticed.
2. Server time confusion. Trades ran across the reset, and a fresh day opened deep underwater.
3. News window ignored. One entry landed two minutes before a rate decision.
4. Two instances running. A forgotten terminal duplicated every order into a second account.
5. Size copied from a personal account. Lots that suited a private balance breached a funded cap.

Every item on that list comes from process rather than prediction. Fix the process, and the strategy finally gets a fair test.

A Pre-Launch Checklist

Work through this before the first live order. It takes an evening and saves a challenge fee.

Check	What good looks like	Warning sign
Daily loss guard	Equity checked on every tick, below the firm’s cap	The program reads balance only
Server time	Reset hour confirmed with support in writing	Local midnight assumed
News filter	Firm’s own tier of releases, with a buffer	Your own idea of important events
Instance count	One terminal, one magic number, one account	An old terminal still logged in
Holding time	Average well above any stated floor	Trades measured in seconds
Position cap	Correlated pairs counted as one exposure	Unlimited simultaneous orders

Tick every row before paying a fee. A blank row on this table usually explains a failed account better than any chart does.

What Nobody Can Tell You in Advance

Some questions have no answer before the fact. Saying so honestly beats guessing.

Nobody can tell you that a program will pass a challenge. Nor can anyone rank robots for prop use, because records rarely share costs, periods or account rules.

Nobody can tell you a firm’s rules will hold either. Terms change, and the change usually arrives by email rather than by negotiation.

What you can establish looks smaller and helps more. You can establish that your code respects each cap, that it stops when told, and that it logs what it did.

FAQ

Do prop firms allow expert advisors at all?

Most allow them, within limits. Typical conditions include a ban on copy trading, a ban on very short holding times, restrictions around major news and a requirement that you own or licensed the code. A minority ban automation entirely, so check the agreement rather than the marketing page before paying any fee.

Can I run the same EA on two funded accounts?

Usually not, at least not with identical settings. Firms treat simultaneous identical fills as duplicate trading, because it concentrates their hedging exposure. If you hold two accounts, ask the firm in writing what separation they expect, and keep the answer.

Does a news filter keep me compliant?

It helps, though only when the list matches theirs. Many traders filter their own idea of important news and still breach a rule on an event the firm named. Copy the firm’s tier of releases into your filter, add a buffer either side, and check the list again whenever the calendar changes.

Which rule ends the most automated accounts?

The daily loss cap, by a wide margin. It measures equity rather than balance, it resets on server time rather than local time, and it triggers during the session rather than at the close. A program that ignores any one of those three details can end an account on a day that finishes green.

Should I test on a demo challenge first?

Yes, and for longer than feels necessary. A demo run with identical caps exposes server time errors, symbol name errors and sizing errors at no cost, and those three faults cause most early failures. Give it several weeks, then step to a paid challenge with the smallest risk setting your rules allow. Results are not guaranteed; past performance is not indicative of future results.